Fix bug in RangeOps::ShiftRight

[EgorBo]

Fixes #123790

For e.g. [0..65535] >> [0..31] I didn't realize I need to swap lower and upper bound for r2, so the result should be:
[0 >> 31 .. 65535 >> 0], not [0 >> 0 .. 65535 >> 31]

No diffs

[dotnet-policy-service]

Tagging subscribers to this area: @JulieLeeMSFT, @dotnet/jit-contrib
See info in area-owners.md if you want to be subscribed.

[Copilot]

Pull request overview

Fixes an incorrect range computation in JIT range analysis for right-shift operations (RangeOps::ShiftRight), addressing #123790 and adding a regression test to prevent recurrence.

Changes:

• Add a JIT regression test for the reported shift/comparison miscompile.
• Rework RangeOps::ShiftRight to compute bounds using opposite shift-count endpoints (lower uses r2.UpperLimit, upper uses r2.LowerLimit) when the shift-count range is constant and within [0..31].
• Remove Range::IsSingleConstValue usage in favor of the existing Range::IsSingleValueConstant.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 2 comments.

File	Description
src/tests/JIT/Regression/JitBlue/Runtime_123790/Runtime_123790.cs	Adds an xUnit regression test for the miscompile scenario from #123790.
src/coreclr/jit/rangecheck.h	Updates range computation logic for ShiftRight and replaces IsSingleConstValue references with IsSingleValueConstant.

[Copilot]

Pull request overview

Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.

[EgorBo]

/azp run Fuzzlyn

[azure-pipelines]

Azure Pipelines successfully started running 1 pipeline(s).

[EgorBo]

PTAL @jakobbotsch @dotnet/jit-contrib fix for ShiftRight (see description). No diffs.

[EgorBo]

/ba-g deadletter

[EgorBo]

/ba-g deadletter